In the post-pandemic era, the rise of Generative Artificial Intelligence (GenAI) has sparked significant concerns from educators and researchers regarding academic integrity. These concerns are highlighted by a 2024 study by Waltzer et al., which illustrates the growing tension between technological innovation and ethical standards in university classrooms. This scoping review systematically maps existing literature to identify how student use of GenAI aligns with documented integrity risks and institutional responses. Following PRISMA-ScR standards, a total of 281 articles were retrieved from Scopus and EBSCOhost between 2023 and 2025, resulting in a finalized dataset of 35 empirical studies. Data extraction utilized the SAMR (Substitution, Augmentation, Modification, Redefinition) model as a diagnostic framework to measure the depth of technology integration. Results indicate a significant demographic concentration, with 88.6% of research situated in Higher Education, and ChatGPT identified as the primary tool utilized. Analysis reveals that while 68.6% of usage falls within the Enhancement phase, there is a critical mismatch between “Process-oriented” risks, such as cognitive erosion, and current “Product-oriented” institutional solutions like detection tools. To address this gap, the study proposes the original SAMR-Integrity-Response (SIR) Framework. This model provides a strategic roadmap for educational institutions, advocating for a shift from defensive regulatory postures at lower integration levels to evolutionary pedagogical pivots, including assessment redesign and process-based grading, at transformative levels. This review equips institutions with the tools to preserve integrity in digitally enabled learning environments.

Keywords: Generative Artificial Intelligence, LLM, Large Language Model, Academic Integrity, Academic Dishonesty